Ridley, Li, and Hill (1998 [this issue]) constructed an intriguing synthesis of multicultural assessment issues merged with what I consider to be the core processes of a scientist-practitioner model of psychological assessment (see Spengler, Strohmer, Dixon, & Shivy, 1995). The authors took on the formidable task of integrating what is thought about and, in some instances, known about multicultural assessment and merged this material with a scientific methodology of hypothesis testing. To paraphrase Luborsky's (1985) initial comment to Gelso and Carter's (1985) major contribution on the relationship: After reading this creative work, I would like to conduct my psychological assessments (and my counseling practice) with Chuck Ridley, Lisa Li, and Carrie Hill. I believe their contribution provides an important, unique, and timely perspective on this topic. Because I mostly agree with their position and recognize that this represents a shared bias, I have attempted to generate alternative perspectives on key issues in the spirit of advancing their creative work and subsequent discussion by others about multicultural assessment.
Ridley and his colleagues (1998) note at the outset that "Psychological assessment should be accurate, thorough, and impartial" (p. 827). This is a lofty goal, which I recognize is a prescription for the ideal, as opposed to a description of what actually occurs, given the limitations of human reasoning (e.g., Nisbett & Ross, 1980), clinical judgment (e.g., Garb, 1998), current use of statistical prediction techniques (e.g., Einhorn, 1986), and psychological tests in relation to nondominant groups. Yet, I would argue that this goal cannot be attained in psychological assessment without some modification and would further argue that this may not be the best goal for psychological assessment. I explain my perspective on this below.
Ridley et al. (1998) emphasize the importance of accurate assessments throughout their discussion. This is an objective with which no one should argue or see reason to do so, but there are at least four unanswered problems related to this objective. First, how should accurate assessment be defined? In the process of synthesizing the extant clinical judgment literature for a series of recta-analyses (e.g., see Spengler, 1998), I and a group of colleagues have identified over 900 empirical clinical judgment studies of mental health judgments made by mental health practitioners or graduate students. However, we can define the accuracy of these judgments in less than only half of the studies and the majority of times, our definition is tentative based on what we consider to be inferential evidence. I would like to have a better idea of what constitutes accurate multicultural assessment outcomes according to the authors. In Spengler et al. (1995), we provide a definition for accurate assessment outcomes that begins with a high level of certainty and ends with a high level of inference. There are specific instances when judgment accuracy can be defined, but only when an acceptable external criterion exists (e.g., neuropsychological findings verified by autopsy, Wedding, 1983; a priori prediction of inpatient violence verified by subsequent client behavior, Werner, Rose, & Yesavage, 1983). At an inferential level, accuracy can be established when, for example, research and practice guidelines exist (e.g., referral to a physician to rule out medical causes before initiating counseling for depression, American Psychiatric Association [APA], 1994) or when client variables (e.g., race, ethnicity) have been shown to bias diagnostic and treatment judgments …